Skip to contentSkip to conversation

driver for i5400 MCH

PATCH RESEND Superseded Mauro Carvalho Chehab 3 participants 5 messages Dec 1, 2008 – Dec 3, 2008

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH RESEND 1/1 driver for i5400 MCH

EDAC driver for i5400 MCH (Seaburg)

This driver adds support for i5400 MCH chipset.

drivers/edac/Kconfig +7 -0
index 5a11e3c..e5d1fcb 100644
@@ -102,6 +102,13 @@ config EDAC_I3000
 	  Support for error detection and correction on the Intel
 	  3000 and 3010 server chipsets.
 
+config EDAC_I5400
+	tristate "Intel 5400 (Seaburg) chipsets"
+	depends on EDAC_MM_EDAC && PCI && X86
+	help
+	  Support for error detection and correction the Intel
+	  i5400 MCH chipset (Seaburg).
+
 config EDAC_I82860
 	tristate "Intel 82860"
 	depends on EDAC_MM_EDAC && PCI && X86_32
drivers/edac/Makefile +1 -0
index e5e9104..4e093e5 100644
@@ -20,6 +20,7 @@ endif
 obj-$(CONFIG_EDAC_AMD76X)		+= amd76x_edac.o
 obj-$(CONFIG_EDAC_I5000)		+= i5000_edac.o
 obj-$(CONFIG_EDAC_I5100)		+= i5100_edac.o
+obj-$(CONFIG_EDAC_I5400)		+= i5400_edac.o
 obj-$(CONFIG_EDAC_E7XXX)		+= e7xxx_edac.o
 obj-$(CONFIG_EDAC_E752X)		+= e752x_edac.o
 obj-$(CONFIG_EDAC_I82443BXGX)		+= i82443bxgx_edac.o
drivers/edac/i5400_edac.c new file +1471 -0

1,471 lines not on this page · raw message

#1 lore raw
Quoted 3 lines · On Mon, 1 Dec 2008 19:46:23 -0200 Mauro Carvalho Chehab…

EDAC driver for i5400 MCH (Seaburg)

This driver adds support for i5400 MCH chipset.

checkpatch has fun with this one.

Quoted 1 line

Both you and Ben Woodard are listed under MODULE_AUTHOR. Should this patch have Ben's Signed-off-by:? And/or his From:?

drivers/edac/i5400_edac.c line 60

Quoted 11 lines

...

drivers/edac/i5400_edac.c new file +8 -0
+/* Device 16,
+ * Function 0: System Address
+ * Function 1: Memory Branch Map, Control, Errors Register
+ * Function 2: FSB Error Registers
+ *
+ * All 3 functions of Device 16 (0,1,2) share the SAME DID
+ */
+#ifndef PCI_DEVICE_ID_INTEL_5400_ERR

This ifndef is unneeded: PCI_DEVICE_ID_INTEL_5400_ERR _is_ defined.

What's more, it is actually bad, because if pci_ids.h were to #define PCI_DEVICE_ID_INTEL_5400_ERR to something other than 0x4030, the ifndef would prevent the compiler from informing us of this potential problem.

drivers/edac/i5400_edac.c

Quoted 12 lines
drivers/edac/i5400_edac.c new file +9 -0
+#define PCI_DEVICE_ID_INTEL_5400_ERR  0x4030	/* Device 16 (0,1,2) */
+#define PCI_DEVICE_ID_INTEL_5400_FBD0 0x4035	/* Device 21 (0,1) */
+#define PCI_DEVICE_ID_INTEL_5400_FBD1 0x4036	/* Device 21 (0,1) */
+#endif
+
 
...
 
+/*
+ * Names to translate bit error into something useful
+ */
+char *error_name[] = {

error_name is a global symbol.

Quoted 27 lines
drivers/edac/i5400_edac.c new file +27 -0
+	[0]  = "Memory Write error on non-redundant retry",
+	[1]  = "Memory or FB-DIMM configuration CRC read error",
+	/* Reserved */
+	[3]  = "Uncorrectable Data ECC on Replay",
+	[4]  = "Aliased Uncorrectable Non-Mirrored Demand Data ECC",
+	/* Unsupported on i5400 */
+	[6]  = "Aliased Uncorrectable Resilver- or Spare-Copy Data ECC",
+	[7]  = "Aliased Uncorrectable Patrol Data ECC",
+	[8]  = "Non-Aliased Uncorrectable Non-Mirrored Demand Data ECC",
+	/* Unsupported */
+	[10] = "Non-Aliased Uncorrectable Resilver- or Spare-Copy Data ECC",
+	[11] = "Non-Aliased Uncorrectable Patrol Data ECC",
+	[12] = "Memory Write error on first attempt",
+	[13] = "FB-DIMM Configuration Write error on first attempt",
+	[14] = "Memory or FB-DIMM configuration CRC read error",
+	[15] = "Channel Failed-Over Occurred",
+	[16] = "Correctable Non-Mirrored Demand Data ECC",
+	/* Unsupported */
+	[18] = "Correctable Resilver- or Spare-Copy Data ECC",
+	[19] = "Correctable Patrol Data ECC",
+	[20] = "FB-DIMM Northbound parity error on FB-DIMM Sync Status",
+	[21] = "SPD protocol Error",
+	[22] = "Non-Redundant Fast Reset Timeout",
+	[23] = "Refresh error",
+	[24] = "Memory Write error on redundant retry",
+	[25] = "Redundant Fast Reset Timeout",
+	[26] = "Correctable Counter Threshold Exceeded",

drivers/edac/i5400_edac.c line 263

Quoted 10 lines
drivers/edac/i5400_edac.c new file +7 -0
+	[27] = "DIMM-Spare Copy Completed",
+	[28] = "DIMM-Isolation Completed",
+};
+
 
...
 
+/* masks for non-fatal error register */
+#define TO_NF_MASK(a)		(((a) & EMASK_M29) | ((a) >> 3))
+#define FROM_NF_FERR(a)		(((a) & EMASK_M29) | (((a) << 3) & ((1 << 30)-1)))

These macros will misbehave if passed an expression with side-effects. Not really a problem as they're only used for constructing constants. I guess if one was being paranoid, one could #undef them again as soon as possible.

Quoted 17 lines
drivers/edac/i5400_edac.c new file +14 -0
+#define FERR_NF_MASK		TO_NF_MASK(ERROR_NF_MASK)
+#define FERR_NF_CORRECTABLE	TO_NF_MASK(ERROR_NF_CORRECTABLE)
+#define FERR_NF_DIMM_SPARE	TO_NF_MASK(ERROR_NF_DIMM_SPARE)
+#define FERR_NF_SPD_PROTOCOL	TO_NF_MASK(ERROR_NF_SPD_PROTOCOL)
+#define FERR_NF_NORTH_CRC	TO_NF_MASK(ERROR_NF_NORTH_CRC)
+#define FERR_NF_RECOVERABLE	TO_NF_MASK(ERROR_NF_RECOVERABLE)
+#define FERR_NF_UNCORRECTABLE	TO_NF_MASK(ERROR_NF_UNCORRECTABLE)
+
 
...
 
+#ifdef CONFIG_EDAC_DEBUG
+/* MTR NUMROW */
+static char *numrow_toString[] = {
+	"8,192 - 13 rows",
+	"16,384 - 14 rows",
+	"32,768 - 15 rows",

drivers/edac/i5400_edac.c line 309

Quoted 10 lines
drivers/edac/i5400_edac.c new file +10 -0
+	"65,536 - 16 rows"
+};
+
+/* MTR NUMCOL */
+static char *numcol_toString[] = {
+	"1,024 - 10 columns",
+	"2,048 - 11 columns",
+	"4,096 - 12 columns",
+	"reserved"
+};

These could possibly be made const. We don't modify them at runtime...

drivers/edac/i5400_edac.c line 519

Quoted 10 lines
drivers/edac/i5400_edac.c new file +7 -0
+#endif
+
 
...
 
+static void i5400_proccess_non_recoverable_info(struct mem_ctl_info *mci,
+				    struct i5400_error_info *info,
+				    unsigned long allErrors)
+{
+	char msg[EDAC_MC_LABEL_LEN + 1 + 90 + 80];

202 bytes. Tolerable, I guess.

Heaven knows why and how this dimension was chosen, and under which circumstances it will overflow :(

It looks like such a hack that perhaps you should give up and use kasprintf()+kfree()?

Quoted 53 lines
drivers/edac/i5400_edac.c new file +53 -0
+	int branch;
+	int channel;
+	int bank;
+	int buf_id;
+	int rank;
+	int rdwr;
+	int ras, cas;
+	int errnum;
+	char *type = NULL;
+
+	if (!allErrors)
+		return;		/* if no error, return now */
+
+	if (allErrors &  ERROR_FAT_MASK)
+		type = "FATAL";
+	else if (allErrors & FERR_NF_UNCORRECTABLE)
+		type = "NON-FATAL uncorrected";
+	else
+		type = "NON-FATAL recoverable";
+
+	/* ONLY ONE of the possible error bits will be set, as per the docs */
+
+	branch = extract_fbdchan_indx(info->ferr_fat_fbd);
+	channel = branch;
+
+	/* Use the NON-Recoverable macros to extract data */
+	bank = nrec_bank(info);
+	rank = nrec_rank(info);
+	buf_id = nrec_buf_id(info);
+	rdwr = nrec_rdwr(info);
+	ras = nrec_ras(info);
+	cas = nrec_cas(info);
+
+	debugf0("\t\tCSROW= %d  Channels= %d,%d  (Branch= %d "
+		"DRAM Bank= %d Buffer ID = %d rdwr= %s ras= %d cas= %d)\n",
+		rank, channel, channel + 1, branch >> 1, bank,
+		buf_id, rdwr_str(rdwr), ras, cas);
+
+	/* Only 1 bit will be on */
+	errnum = find_first_bit(&allErrors, ARRAY_SIZE(error_name));
+
+	/* Form out message */
+	snprintf(msg, sizeof(msg),
+		 "%s (Branch=%d DRAM-Bank=%d Buffer ID = %d RDWR=%s RAS=%d CAS=%d "
+		 "%s Err=0x%lx (%s))",
+		 type, branch >> 1, bank, buf_id, rdwr_str(rdwr), ras, cas, type,
+		 allErrors, error_name[errnum]);
+
+	/* Call the helper to output message */
+	edac_mc_handle_fbd_ue(mci, rank, channel, channel + 1, msg);
+}
+
+/*

drivers/edac/i5400_edac.c line 582

Quoted 10 lines
drivers/edac/i5400_edac.c new file +10 -0
+ * i5400_process_fatal_error_info(struct mem_ctl_info *mci,
+ * 				struct i5400_error_info *info,
+ * 				int handle_errors);
+ *
+ *	handle the Intel NON-FATAL errors, if any
+ */
+static void i5400_process_nonfatal_error_info(struct mem_ctl_info *mci,
+					struct i5400_error_info *info)
+{
+	char msg[EDAC_MC_LABEL_LEN + 1 + 90 + 80];

dittoes.

Quoted 83 lines
drivers/edac/i5400_edac.c new file +80 -0
+	unsigned long allErrors;
+	int branch;
+	int channel;
+	int bank;
+	int rank;
+	int rdwr;
+	int ras, cas;
+	int errnum;
+
+	/* mask off the Error bits that are possible */
+	allErrors = FROM_NF_FERR(info->ferr_nf_fbd & FERR_NF_MASK);
+	if (!allErrors)
+		return;		/* if no error, return now */
+
+	/* ONLY ONE of the possible error bits will be set, as per the docs */
+
+	if (allErrors & (ERROR_NF_UNCORRECTABLE | ERROR_NF_RECOVERABLE)) {
+		i5400_proccess_non_recoverable_info(mci, info, allErrors);
+		return;
+	}
+
+	/* Correctable errors */
+	if (allErrors & ERROR_NF_CORRECTABLE) {
+		debugf0("\tCorrected bits= 0x%lx\n", allErrors);
+
+		branch = extract_fbdchan_indx(info->ferr_nf_fbd);
+
+		channel = 0;
+		if (REC_ECC_LOCATOR_ODD(info->redmemb))
+			channel = 1;
+
+		/* Convert channel to be based from zero, instead of
+		 * from branch base of 0 */
+		channel += branch;
+
+		bank = rec_bank(info);
+		rank = rec_rank(info);
+		rdwr = rec_rdwr(info);
+		ras = rec_ras(info);
+		cas = rec_cas(info);
+
+		/* Only 1 bit will be on */
+		errnum = find_first_bit(&allErrors, ARRAY_SIZE(error_name));
+
+		debugf0("\t\tCSROW= %d Channel= %d  (Branch %d "
+			"DRAM Bank= %d rdwr= %s ras= %d cas= %d)\n",
+			rank, channel, branch >> 1, bank,
+			rdwr_str(rdwr), ras, cas);
+
+		/* Form out message */
+		snprintf(msg, sizeof(msg),
+			 "Corrected error (Branch=%d DRAM-Bank=%d RDWR=%s RAS=%d "
+			 "CAS=%d, CE Err=0x%lx (%s))", branch >> 1, bank,
+			 rdwr_str(rdwr), ras, cas, allErrors,
+			error_name[errnum]);
+
+		/* Call the helper to output message */
+		edac_mc_handle_fbd_ce(mci, rank, channel, msg);
+
+		return;
+	}
+
+	/* Miscelaneous errors */
+	errnum = find_first_bit(&allErrors, ARRAY_SIZE(error_name));
+
+	branch = extract_fbdchan_indx(info->ferr_nf_fbd);
+
+	i5400_mc_printk(mci, KERN_EMERG,
+			"Non-Fatal misc error (Branch=%d Err=%#lx (%s))",
+			branch >> 1, allErrors, error_name[errnum]);
+}
+
 
...
 
+static void i5400_put_devices(struct mem_ctl_info *mci)
+{
+	struct i5400_pvt *pvt;
+
+	pvt = mci->pvt_info;
+
+	/* Decrement usage count for devices */
+	if (pvt->branch_1)

drivers/edac/i5400_edac.c line 717

Quoted 10 lines
drivers/edac/i5400_edac.c new file +10 -0
+		pci_dev_put(pvt->branch_1);
+
+	if (pvt->branch_0)
+		pci_dev_put(pvt->branch_0);
+
+	if (pvt->fsb_error_regs)
+		pci_dev_put(pvt->fsb_error_regs);
+
+	if (pvt->branchmap_werrors)
+		pci_dev_put(pvt->branchmap_werrors);

pci_dev_put(NULL) is legal.

Quoted 61 lines
drivers/edac/i5400_edac.c new file +58 -0
+}
+
 
...
 
+static int i5400_probe1(struct pci_dev *pdev, int dev_idx)
+{
+	struct mem_ctl_info *mci;
+	struct i5400_pvt *pvt;
+	int num_channels;
+	int num_dimms_per_channel;
+	int num_csrows;
+
+	debugf0("MC: " __FILE__ ": %s(), pdev bus %u dev=0x%x fn=0x%x\n",
+		__func__,
+		pdev->bus->number,
+		PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
+
+	/* We only are looking for func 0 of the set */
+	if (PCI_FUNC(pdev->devfn) != 0)
+		return -ENODEV;
+
+	/* Ask the devices for the number of CSROWS and CHANNELS so
+	 * that we can calculate the memory resources, etc
+	 *
+	 * The Chipset will report what it can handle which will be greater
+	 * or equal to what the motherboard manufacturer will implement.
+	 *
+	 * As we don't have a motherboard identification routine to determine
+	 * actual number of slots/dimms per channel, we thus utilize the
+	 * resource as specified by the chipset. Thus, we might have
+	 * have more DIMMs per channel than actually on the mobo, but this
+	 * allows the driver to support upto the chipset max, without
+	 * some fancy mobo determination.
+	 */
+	i5400_get_dimm_and_channel_counts(pdev, &num_dimms_per_channel,
+					&num_channels);
+	num_csrows = num_dimms_per_channel * 2;
+
+	debugf0("MC: %s(): Number of - Channels= %d  DIMMS= %d  CSROWS= %d\n",
+		__func__, num_channels, num_dimms_per_channel, num_csrows);
+
+	/* allocate a new MC control structure */
+	mci = edac_mc_alloc(sizeof(*pvt), num_csrows, num_channels, 0);
+
+	if (mci == NULL)
+		return -ENOMEM;
+
+	debugf0("MC: " __FILE__ ": %s(): mci = %p\n", __func__, mci);
+
+	mci->dev = &pdev->dev;	/* record ptr  to the generic device */
+
+	pvt = mci->pvt_info;
+	pvt->system_address = pdev;	/* Record this device in our private */
+	pvt->maxch = num_channels;
+	pvt->maxdimmperch = num_dimms_per_channel;
+
+	/* 'get' the pci devices we want to reserve for our use */
+	if (i5400_get_devices(mci, dev_idx))
+		goto fail0;
+

drivers/edac/i5400_edac.c line 1311

Quoted 10 lines
drivers/edac/i5400_edac.c new file +10 -0
+	/* Time to get serious */
+	i5400_get_mc_regs(mci);	/* retrieve the hardware registers */
+
+	mci->mc_idx = 0;
+	mci->mtype_cap = MEM_FLAG_FB_DDR2;
+	mci->edac_ctl_cap = EDAC_FLAG_NONE;
+	mci->edac_cap = EDAC_FLAG_NONE;
+	mci->mod_name = "i5400_edac.c";
+	mci->mod_ver = I5400_REVISION;
+	mci->ctl_name = i5400_devs[dev_idx].ctl_name;

This is a bit strange. i5400_devs[] has a single entry, so what's the point in indexing it with `dev_idx', which MUST have a value of zero anyway?

Quoted 52 lines
drivers/edac/i5400_edac.c new file +52 -0
+	mci->dev_name = pci_name(pdev);
+	mci->ctl_page_to_phys = NULL;
+
+	/* Set the function pointer to an actual operation function */
+	mci->edac_check = i5400_check_error;
+
+	/* initialize the MC control structure 'csrows' table
+	 * with the mapping and control information */
+	if (i5400_init_csrows(mci)) {
+		debugf0("MC: Setting mci->edac_cap to EDAC_FLAG_NONE\n"
+			"    because i5400_init_csrows() returned nonzero "
+			"value\n");
+		mci->edac_cap = EDAC_FLAG_NONE;	/* no csrows found */
+	} else {
+		debugf1("MC: Enable error reporting now\n");
+		i5400_enable_error_reporting(mci);
+	}
+
+	/* add this new MC control structure to EDAC's list of MCs */
+	if (edac_mc_add_mc(mci)) {
+		debugf0("MC: " __FILE__
+			": %s(): failed edac_mc_add_mc()\n", __func__);
+		/* FIXME: perhaps some code should go here that disables error
+		 * reporting if we just enabled it
+		 */
+		goto fail1;
+	}
+
+	i5400_clear_error(mci);
+
+	/* allocating generic PCI control info */
+	i5400_pci = edac_pci_create_generic_ctl(&pdev->dev, EDAC_MOD_STR);
+	if (!i5400_pci) {
+		printk(KERN_WARNING
+			"%s(): Unable to create PCI control\n",
+			__func__);
+		printk(KERN_WARNING
+			"%s(): PCI error report via EDAC not setup\n",
+			__func__);
+	}
+
+	return 0;
+
+	/* Error exit unwinding stack */
+fail1:
+
+	i5400_put_devices(mci);
+
+fail0:
+	edac_mc_free(mci);
+	return -ENODEV;
+}

drivers/edac/i5400_edac.c line 1432

Quoted 10 lines
drivers/edac/i5400_edac.c new file +7 -0
+
 
...
 
+static struct pci_driver i5400_driver = {
+	.name = KBUILD_BASENAME,
+	.probe = i5400_init_one,
+	.remove = __devexit_p(i5400_remove_one),
+	.id_table = i5400_pci_tbl,
+};

Only one other driver in the tree uses KBUILD_BASENAME, and that is drivers/edac/i5100_edac.c, from which I suspect the above was copied.

I suspect that something is being done wrongly here.

Quoted 3 lines

...

#2 lore raw
Quoted 7 lines · On Tue, 2 Dec 2008 22:33:47 -0800 Andrew Morton <[email protected]>…
Quoted 3 lines · On Mon, 1 Dec 2008 19:46:23 -0200 Mauro Carvalho Chehab…

EDAC driver for i5400 MCH (Seaburg)

This driver adds support for i5400 MCH chipset.

checkpatch has fun with this one.

Ok. I fixed the remaining errors.

I'll just keep the 80 char warns on lines with comments explaining each field, like:

WARNING: line over 80 characters
#200: FILE: drivers/edac/i5400_edac.c:151:
+ EMASK_M9 = 1<<8, /* Non-Aliased Uncorrectable Non-Mirrored Demand Data ECC */

Where breaking the line wouldn't be that fun.

Quoted 4 lines
Quoted 1 line

Both you and Ben Woodard are listed under MODULE_AUTHOR. Should this patch have Ben's Signed-off-by:? And/or his From:?

I'll ask his SOB.

Quoted 7 lines
Quoted 1 line

+#ifndef PCI_DEVICE_ID_INTEL_5400_ERR

This ifndef is unneeded: PCI_DEVICE_ID_INTEL_5400_ERR _is_ defined.

What's more, it is actually bad, because if pci_ids.h were to #define PCI_DEVICE_ID_INTEL_5400_ERR to something other than 0x4030, the ifndef would prevent the compiler from informing us of this potential problem.

Fixed.

Quoted 3 lines
Quoted 1 line

+char *error_name[] = {

error_name is a global symbol.

Changed. Also changed it to const.

Quoted 8 lines

drivers/edac/i5400_edac.c line 263

Quoted 3 lines
drivers/edac/i5400_edac.c new file +3 -0
+/* masks for non-fatal error register */
+#define TO_NF_MASK(a)		(((a) & EMASK_M29) | ((a) >> 3))
+#define FROM_NF_FERR(a)		(((a) & EMASK_M29) | (((a) << 3) & ((1 << 30)-1)))

These macros will misbehave if passed an expression with side-effects. Not really a problem as they're only used for constructing constants. I guess if one was being paranoid, one could #undef them again as soon as possible.

Good point. I've converted both macros into inline functions, since gcc will likely optimize the code into a constant, instead of calculating at runtime.

Quoted 10 lines
Quoted 10 lines
drivers/edac/i5400_edac.c new file +10 -0
+#ifdef CONFIG_EDAC_DEBUG
+/* MTR NUMROW */
+static char *numrow_toString[] = {
+	"8,192 - 13 rows",
+	"16,384 - 14 rows",
+	"32,768 - 15 rows",
+	"65,536 - 16 rows"
+};
+
+/* MTR NUMCOL */
Quoted 8 lines

drivers/edac/i5400_edac.c line 309

Quoted 6 lines
drivers/edac/i5400_edac.c new file +6 -0
+static char *numcol_toString[] = {
+	"1,024 - 10 columns",
+	"2,048 - 11 columns",
+	"4,096 - 12 columns",
+	"reserved"
+};

These could possibly be made const. We don't modify them at runtime...

Fixed.

Quoted 7 lines

202 bytes. Tolerable, I guess.

Heaven knows why and how this dimension was chosen, and under which circumstances it will overflow :(

It looks like such a hack that perhaps you should give up and use kasprintf()+kfree()?

I also didn't like that code. It is present also on other edac drivers.

I'll fix this on a later patch. I'll probably propose some patches to the other edac drivers that use this kind of construction.

Quoted 6 lines
Quoted 6 lines
drivers/edac/i5400_edac.c new file +6 -0
+	/* Decrement usage count for devices */
+	if (pvt->branch_1)
+		pci_dev_put(pvt->branch_1);
+
+	if (pvt->branch_0)
+		pci_dev_put(pvt->branch_0);
Quoted 8 lines

drivers/edac/i5400_edac.c line 717

Quoted 6 lines
drivers/edac/i5400_edac.c new file +6 -0
+
+	if (pvt->fsb_error_regs)
+		pci_dev_put(pvt->fsb_error_regs);
+
+	if (pvt->branchmap_werrors)
+		pci_dev_put(pvt->branchmap_werrors);

pci_dev_put(NULL) is legal.

Fixed.

Quoted 5 lines

drivers/edac/i5400_edac.c line 1311

Quoted 1 line
drivers/edac/i5400_edac.c new file +1 -0
+	mci->ctl_name = i5400_devs[dev_idx].ctl_name;

This is a bit strange. i5400_devs[] has a single entry, so what's the point in indexing it with `dev_idx', which MUST have a value of zero anyway?

I'm not sure if we covered all PCI ID's for i5400 (maybe some i5400 variant may be released?). That's why I preferred to keep a table: makes easier to add other entries there, but I see your point.

I'll add a code to prevent dev_idx values equal or higher than ARRAY_SIZE, to avoid the risk of trying to acess something outside the buffer.

Quoted 11 lines

drivers/edac/i5400_edac.c line 1432

Quoted 6 lines
drivers/edac/i5400_edac.c new file +6 -0
+static struct pci_driver i5400_driver = {
+	.name = KBUILD_BASENAME,
+	.probe = i5400_init_one,
+	.remove = __devexit_p(i5400_remove_one),
+	.id_table = i5400_pci_tbl,
+};

Only one other driver in the tree uses KBUILD_BASENAME, and that is drivers/edac/i5100_edac.c, from which I suspect the above was copied.

I suspect that something is being done wrongly here.

Fixed.

I'll run some tests here, and submit a version 2 of the patch soon.

Cheers,
Mauro

#3 lore raw
In reply to Andrew Morton

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Quoted 5 lines · Andrew Morton
Quoted 2 lines

Both you and Ben Woodard are listed under MODULE_AUTHOR. Should this patch have Ben's Signed-off-by:? And/or his From:?

Go ahead and add a:

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Red Hat - http://enigmail.mozdev.org

iD8DBQFJNufYYuGAPpw3tAARAlA5AJ4hNh6Mzzs60atqL06MkoD7RulsbwCdGSP7 DIQiX8F4+1eQXYeZPSQkc0M=
=anBc
-----END PGP SIGNATURE-----

#4 lore raw

PATCH v2 1/1 edac: driver for i5400 MCH (Seaburg)

From: Mauro Carvalho Chehab <[email protected]>

This patch adds a driver for support errors reported by i5400 MCH chipset (Seaburg)

drivers/edac/Kconfig +7 -0
index e0dbd38..c1fcbed 100644
@@ -109,6 +109,13 @@ config EDAC_X38
 	  Support for error detection and correction on the Intel
 	  X38 server chipsets.
 
+config EDAC_I5400
+	tristate "Intel 5400 (Seaburg) chipsets"
+	depends on EDAC_MM_EDAC && PCI && X86
+	help
+	  Support for error detection and correction the Intel
+	  i5400 MCH chipset (Seaburg).
+
 config EDAC_I82860
 	tristate "Intel 82860"
 	depends on EDAC_MM_EDAC && PCI && X86_32
drivers/edac/Makefile +1 -0
index 62c2d9b..b751969 100644
@@ -20,6 +20,7 @@ endif
 obj-$(CONFIG_EDAC_AMD76X)		+= amd76x_edac.o
 obj-$(CONFIG_EDAC_I5000)		+= i5000_edac.o
 obj-$(CONFIG_EDAC_I5100)		+= i5100_edac.o
+obj-$(CONFIG_EDAC_I5400)		+= i5400_edac.o
 obj-$(CONFIG_EDAC_E7XXX)		+= e7xxx_edac.o
 obj-$(CONFIG_EDAC_E752X)		+= e752x_edac.o
 obj-$(CONFIG_EDAC_I82443BXGX)		+= i82443bxgx_edac.o
drivers/edac/i5400_edac.c new file +1476 -0

1,476 lines not on this page · raw message